Depression Adolescence

8/13/2019 Depression Adolescence

1/12


2/12

134 LEWIN SOHN, HOPS, ROBERTS, SEELEY, AN DR EW Spervasive diagnosisif theymetrevised DSM-III(DSM-IH-R;American Psy chiatric Association, 19 87) criteria from morethano neinformant ,a situational diagnosis iftheym etcrite-ria from only one inform an t, or a weak, pervasive diagnosisbycom bining symptoms reported by allinformants.Th en u m-ber ofsubjects also varied markedlyand sample sizes wereof -ten toosmalltoyield stable resultsand topermitfine-grainedcomparisons between age and gender groups. Thus, the Ka-shani,Carlson, et al. (1987)prevalence of 8% was based on asamp le sizeof150,and the Birdet al.(1988) samplewasbasedon 777 children (aged4-16),386 of whom were interviewed. Itis also important to note that none of these researchers pre-sented information about the incidenceofdepression (i.e.,therateofnewcasesofdepressionin thepopulation duringaspeci-fied period of time) fo rwhich a prospective design is needed.Similarly, none of the researchers provided basic informationon things such as the duration of episodes and multiple epi-sodes.In this article we report findings from the Oregon Adoles-cent Depression Project (OADP), a large-scale, community-based investigation of the epidemiology of depression an dother psychiatric disorders among a high school population.The d istinguish ing characteristics of the OADP are as follows:(a) a community-based epidemiological survey; (b) a longitu-dinal ,prospective design, withfollow-up assessments reportedafter 1year; (c) the use of contemporary diagnostic criteria(DSM-III-R); and (d) the inclusionofdiagnostic inform ationencompassinga broad rangeof psychiatric disorders. The re-search satisfies the criteria outlined by A ngst, Dobler-Mikola,an d Binder(1984),Flemingan d Offord (1990),HirschfeldandCross(1982),and M itchell, McCauley, Burke,and Moss(1988)forgen erating mo re reliable and valid data on the epidemiologyofaffective and other psychiatric disorders.Wepresent data on the prevalence (point and lifetime), inci-dence(1year),and comorbidity foraffective and selected otherpsychiatric disorders in middle-to-older (14-18) adolescents.Wealso present inform ation on other epidemiological charac-teristics such as severity, onset age, and d uration , a s well as theassociation of the occurrence of depression with age andgender.Given that adolescents had significantly more depres-sive disorders than children in studies that included both agegroups (B ird et al., 1988; Flem ing,Offord, & Boyle, 1989; Rut-ter, Tizard, Yule, Graham, & Whitmore, 1976)and that thepercentageof deaths attributable to suicide increased sharplybetween 15 and 19 years of age (National Center for HealthStatistics, 1989), we hypothesized an increase in depressionduring adolescence. Furthermore, given that there was nogenderdifference at the youngest ages (Fleming et al., 1989;Kashani et al.,1983;Velezet al.,1989)and that by adulthoodthere was a 2:1 female preponderance (Weissman & Myers,1978), we also hypothesized an interaction between age andgender such that prevalence rates fo r female subjects wouldincrease disproportionately with increasing age.

MethodOverview

Adiagnostic interview was conducted with each adolescent at po intof entry into the study (Time 1). Immediately prior to the interview,

subjects completed a questio nna ire that inclu ded measures of psycho-socialconstructs (Hops, Lewinsohn, Andrews,& Roberts, 1990).Ascreening instrument was also completed twice, once as part of thequestionnairea nd again approximately1week later. Parents providedinformation regarding their education and occupation via short mailquest ionnaires. A pproximately 1 year after their initial assessment(Time 2),participants were reassessed via interviewsan dquest ion-naires. Written informed consent wa s obtained from al l adolescentpart icipantsand from parentsorguardians. Ea ch participan t received$25 and had $5placed in to their schoolfund fo reach assessment. Thissample hasb een previously described in Lewinsohn , Rohde,Seeley,and Hops(1991);Rohde, Lewin sohn, and Seeley(1991);and Roberts,Lewinsohn, andSeeley(1991).

SubjectsThe po pulation for this study was the total enro llmen t (approxi-mately10,200)of nin e high schools(Grades9-12)in two urban com-munities (metropolitan populations of approximately 200,000) an dthree rura l com mu nities in west central Oregon. Schools were chosenbecause of their location(within100 miles of the project). All of the 10

schools we approached agreed toparticipate,although 1 declined later.Three cohorts were recruited in1987,1988,and 1989 and consistedof 352,864,and 4 94 students, respectively. Thetotalcompleted Time 1sample size was1,710 and the completed Time 1-Time 2 panel was1,508.Th e followingsam pling strategyw asused.1 . At the beginning ofeach academicyear, parents of all (includin gthose inspecial classes) students enrolled ineachof the participatingschools were sentaletter describing theproposedresearch an daskingfor permission fo r their offspring to be included in the potentialsample.2. Students whose paren tsdid notreturnthe decline card (passiveconsent procedure) constituted the sampling frame. The proportiondeclining at this stage ranged from 4%(Cohort 3) to 8%(Cohorts1and 2).3. Sampling fractionso f 10%,18.5%, and 20% were used foreach

cohort, a nd sam pling within each school was proportional to the sizeofthe school, the size of the grade within school, and the gender withinthe grade.The selected students and their parents received two letters, the firstwelcoming them to thestudyand asecond inform ing them that theywould receivea phone call within the next fewdays. Th e callerat -tempted to schedule the adolescent for an interview. Participants with-out phones were sent a note asking them to contact the institute. Ifthere was no response, a m ember of thestaffw as sent to the student'shomeo rschool toexplainthe projectand toschedulea n interview.

Response RatesBecause parental involvement in the diagnostic interview was re-quired in Cohort 1, the decline rate among those for whom passiveconsent had been obtained earlier was relatively high (48%).This re-quirement w as dropped forCohorts 2 and 3, and the respective declinerates dropped to 38% and32%, respectively, resulting in an overallparticipationra te of 61%. Toassessdifferences between participan tsan ddecliners,weobtainedbrief demographic inform ation on key vari-ables from the latter by telephone, including reasonsfor decline. Inmostcases,adolescents expresseddisinterest;12%wereoverruled bytheir parents,12% thought the assessment was too personal, and 4%providedvariou s other explanations, including beingtooshy,toobusy,and soforth.W efoundasignificant bu tsmall relationbetween gradeleveland the reason for decline,x2(9 ,N= 938)=21.4,p


3/12

ADOLESCENT PSYCHOPATHOLOGY 135older studentsweremorelikelytodeclinebecausethey werenotinter-ested.

school,35.5%withpa rtial college education, and 20.3% with an aca-demic or pro fessional degree.

Representativeness of the SampleSeveralchecks on the representativeness of the sam ple were m ade.

First, we compared the demographic characteristics of the samplewiththe 1980 census a ndfoundn odifferenceson gender, ethnic status,or paren tal edu cation level. Not surprisingly, our sam ple had signifi-cantly mo re children u nder 18 years of age in the home and a slightlyhigher proportion of two-parent fam ilies. Second, we compa red ourparticipants with those who declined on demographic informationobtained fromdecliners by telephone.Differenceswere minimal. Fam-ilies were similar on gender of head of household, family size, andnum ber of parents in the household. Although the decliners' meansocioeconomic status (SES) wassignificantly lower than that of theparticipants,F(\, 2023)=97.0,p


4/12

136 L E W I N S O H N , HOPS, ROBERTS, SEELEY, AN DRE W Spresenta tT i me1, inadd i t ionto theonsetof newdisorders, byask ingrespondents to provide detailedinformat ionabou t their men tal s ta tuss ince Tim e 1. Usingcareful probing, th e interviewer establishes th eapproxim ate datesf orcritical transition points (i.e.,offset of aprev iousepisode, onset of a new episode). On the basis of this inform ation , theinterviewersmade weeklyratingsofsymptom levelsforeach diagno sisassigned durin gtheinitia l (Time 1)inte rv iewfo reach ne wepisodeo fdisorder. These ratings were madeon a6-point scale that ind icated,foreach week,if thepartic ipant(a) hadcont inuedtomee tth e DSM-IU-Rcriteria for the index disorder (5 or 6); (b) had marked or moderatesympto ms witho ut meeting cr iteria (4);(c)experienced par tia l rem is-sion (3);(d) was in aresidual state (2);or (e) had no residual symptoms(1).A n exis ting episode was assumed to have ended by symptom rat-ings of 1 or 2 for 8 consecutive weeks. Th e interviewers a lso madesymptomseve rity ratingsfor allepisodesofdisorder s ince Tim e1.T heinterrater reliability of these diagnoses was high and comparable tothat reportedb y theLIFE'S authors (K el leretal.,1987).Becauseof thelo w frequen cies of m any disorders , we com puted kappas across a l ldisorders. The kappas for any versus no disorder at T i me 2 and be-tween T i me s1 and 2were.87 and .72, respectively.

By prov id ing a r igorous definition of recovery, th e SADS-LIFEmethodology al lowed for the rela tively precise determ inatio n of thedurationo fepisodes ofdisorder that existeda tT i me 1. TheL IFE alsoprobed for the occurrence ofn ewdisorders (i.e., since Tim e1),the dateofonse tof newepisodes ofdisorder, an d intervals between differentepisodes of disorder. At the Time 2 interview, the interviewers alsoelicited information aboutand rated the presence andseverityof de-pression symptom s and completed the Ham il ton R ating Scale for De-press ion (Ham il ton,1960) forcur rent symptom san dw orst past symp-t oms in the Time 1-Time2inte rva l .Th einterviewersalso m adeAxisVGlobal Assessmen tofFunc t ion-in g(GAF;DSM-IH-R, Am erica n Psychiatr ic Association, 1987) ra t-ingsfor thecur rent levelo f funct ioning (i.e.,at the T i me 2evaluation)and the highestleveloffunct ion ingdur ingtheperiod between Times1an d 2. The interviewersfollowed th e DSM-IH-R direction to givea noveralljud gm ent of the adolescent'spsychological,social, and occupa-t iona l(academic)functioningand need for treatm ent. Thus, the GAF isa global judgm ent that combines sym ptom severity, social role func-t ioning, and perceived need fo r treatment. Th e scale ranges from 1 to90. Scores of 81-90 represent very good fun ctionin g in a l l areas,whereas1-10 represents severely impaired fun ctionin g.

Ratings byClinical Child PsychiatristsTo prov ide informa t ion about th e cl inical s ignificanceof episodesof DSM-IH-R affective disorders detected by the K-SADS methodol-ogy,three childpsychiatrists2rated all subjects with a current diagnosisof affective disorder at T i me 1 n =50) and those wh o developed an

episode of affective disorder betweenTimes 1 and 2( n= 121)for (a) cu rrent level of functionin g, (b) highest level offunctioningd u r i ngthe pastyear, (c) severity ofdepression, and (d) needfo rtreatment.These ratings weredone onspecially developedandcarefully anchored6- or 7-point scales for each of the dimen sions. The p sychiatrists wereinstructedto use theadolescent p atients from their clinicalpracticesasa f rameof reference. Ourscaleswerecloselymodeled after existingscales (e.g., the Children's Global Assessment Scale; Shaffer et al.,1983).Theratings were madeon thebasisof the intakenotes writtenbytheK-SADS interviewers. These notes(two-threepages) includedalisting of all cur rent and past diagnoses and symptoms; onset age;episode duration;and as umma r yof thesubjects'educational, social,and heal th his tory and of the family environment. About one thirdwererated bymore thanonepsychiatristand interraterreliabilitiesaveraged across raterpairsranged from .88 (fo r severity) to .58 (fo rneed fo rtreatment) .

Other MeasuresTh e Center for Epidemic/logic Studies-Dep ression Scale (Radloff,

1977). The CES-D was includ ed as a self-report me asure of depres-sivesym ptom atology. This 20-item scale assesses the occurre nce ofdepressivesym ptom s du ring the past week. The scale has been shownto have adequate psychometr ic properties on an adolescent sample(Roberts,A ndrews, Lewinsohn,&Hops,1990).Th e Beck DepressionInventory (Beck, 1967; Beck, Ward, Mendelson,Mock, Erbaugh, 1961). Th e BDI, another self-report measure ofdepression, was a lso include d. This21-item scale assesses th e pr esen ceo fdepressive symp tom s durin gth epast weekand hasgood psychomet-ri cproperties when usedwithadolescent samples (Teri,1982;Rober tsetal., 1990,1991).

ResultsPointand Lifetime PrevalenceofDSM-HI-R Disorders

Tables 1 and 2 show the pointan d lifetime prevalence bygender, respectively,fo reachof the majorDSM -III-R diagnos-tic categoriesat Times 1 and 2. In interpreting the resultsofTables1 an d 2 , it isimportanttokeepinm ind thatthepointan dlifetime prevalences fo r specific disorders included subjectswh oh admorethan one disorder. Ofthosewitha lifetimehis-toryof amentaldisorder(n=634),31%reportedhavingexperi-encedanothermental disorder;amongtheadolescentswithalifetimehistoryofunipolardepression(n =348),42 %reportedhavingexperienced another mentaldisorder;and a mo ng thosewithacurren t(a tTime1 )diagnosisofunipolardepression(n =50), 66% had a history ofano the rmental disorder,and 34%reportedhavingexperiencedapreviousepisodeofdepression.Thedegreeofcomorbidity(overthe lifetime)betweena llof themajor DSM-III-R disordersisshown in Table3. As can beseen,therewas substantial co mo rbiditybetweenall o f thedis-orders except for adjustment disorder, which wascomorbidonlywith substancedependencean dabuse.Adiagnosisof ad-justmentdisorder madeit less likely thatsubjectswould haveha ddiagnosesofunipolardepression,disruptive b ehaviordis-orders,bipolar disorders,or eatingdisorders.Alikelyexplana-tion isthat, as per the DSM-III-R, meeting criteriafor anyspecific mental disorder is an exclusion criterionf or adjust-mentdisorder.Comparisons of the Time1 and Time 2 pointprevalenceratesfor the Tim e 1-Time2panel(n =1,508),b ymeansof thecriticalratioz,indicatedthatatT im e2,when the participantswere 1yearolder,thereha dbeenan increasein thepointpreva-lence of alcoholdependenceandabuse(z =2.60,p


5/12

ADOLESCENT PSYCHOPATHOLOGY 137Table1Point PrevalenceRates P P R s ) o/DSM-III-RDisordersby Gender

Female*Psychiatric diagnosis

Unipolar depressionMajor depressionDysthymiaBipolar disorderAnxietydisordersPanicAgoraphobiaSocial phobiaSimplephobiaObsessive-compulsiveSeparation anxietyOveranxiousDisruptive behaviordisorderAttention-deficithyperactivityConductOppositional defiantSubstance use disordersAlcoholdependenceandabuseDrug dependence and abuseCannabisHard drugsCocaineAmphetaminesEating disordersAnorexianervosaBulimia nervosaAdjustment disorderOther disordersAny diagnosis

PPR3.823.370.560.454.710.450.671.572.020.110.340.671.010.340.340.452.131.121.571.350.340.000.220.340.000.341.120.0011.22

SE0.64*0.60*0.250.220.71***0.220.270.42**0.47*0.110.190.270.34**0.190.190.22**0.480.350.420.390.190.000.160.190.000.190.35*0.001.06*

Time1 (%) Time2(% )Male

PP R1.951.710.490.121.470.240.120.240.730.000.000.242.690.490.851.472.560.852.082.080.490.120.240.000.000.000.240.127.81

SE0.480.450.240.120.420.170.120.170.300.000.000.170.570.240.320.420.550.320.500.500.240.120.170.000.000.000.170.120.94

Total'PPR2.922.570.530.293.160.350.410.941.400.060.180.471.810.410.130.942.340.991.811.700.410.060.230.180.000.180.700.069.59

Note. DSM-IH-R =DiagnosticandStatistical Manualof Mental Disorders(rev.a = 8 9 1 . b= 819. c = l , 7 1 0 . = 81I3. en= 69:8. f= 1,508.

SE0.410.380.180.130.420.140.150.230.280.060.100.170.320.150.090.230.370.240.320.310.150.060.120.100.000.100.200.060.72

3rded.).

FemaledPPR3.703.580.250.251.980.620.120.370.620.000.120.120.250.000.120.121.601.110.620.490.120.000.120.490.000.491.360.008.02

SE0.660.650.170.170.49*0.280.120.210.280.000.120.120.170.000.120.120.44**0.370.28**0.25**0.120.000.120.250.000.250.41*0.000.96

MalecPPR2.582.580.000.140.570.000.000.000.430.000.000.140.720.140.140.573.872.152.292.150.290.140.000.000.000.000.290.147.76

SE0.600.600.000.140.290.000.000.000.250.000.000.140.320.140.140.290.730.550.570.550.200.140.000.000.000.000.200.140.69

TotalfPP R SE3.18 0.453.12 0.450.13 0.090.20 0.111.33 0.290.33 0.150.07 0.070.20 0.110.53 0.190.00 0.000.07 0.070.13 0.090.46 0.180.07 0.070.13 0.090.33 0.152.65 0.411.59 0.321.39 0.301.26 0.290.20 0.110.07 0.070.07 0.070.27 0.130.00 0.000.27 0.130.86 0.240.07 0.077.82 0.69

*p


6/12

138 LEW I NSO H N, HOPS, ROBERTS, SEELEY,A N D R E W STable2Lifetime PrevalenceRates(LPRs) o/DSM-III-R DisordersbyGender

Female3Psychiatricdiagnosis

UnipolardepressionMajordepressionDysthymiaBipolardisorderAnxietydisordersPanicAgoraphobiaSocial phobiaSimplephobiaObsessive-compulsiveSeparationanxietyOveranxiousDisruptivebehavior disorderAttention-deficithyperactivityConductOppositional defiantSubstanceu sedisordersAlcoholDrugsCannabisHard drugsCocaineAmphetaminesEatingdisordersAnorexia nervosaBulimianervosaAdjustment disorderOther disordersAn ydiagnosis

LP R27.1624.804.040.5611.671.121.122.362.810.345.841.804.71

1.801.681.808.194.835.844.262.690.341.801.350.450.907.070.3442.09

SE1.49***1.45***0.66*0.251.08***0.350.35*0.51**0.55*0.190.79***0.450.71***

0.45**0.43***0.450.920.720.790.68*0.540.190.450.39**0.220.32**0.860.191.65***

Time1 (%)Male

LP R12.9411.602.320.615.620.490.240.491.100.732.440.7310.134.524.883.178.424.276.726.592.440.371.220.120.000.125.250.2431.62

SE1.171.120.530.270.810.240.170.240.360.300.540.301.06

0.730.750.610.970.710.880.870.540.210.380.120.000.120.780.171.63

Total0LP R20.3518.483.220.588.770.820.701.461.990.534.211.297.31

3.103.222.468.304.566.265.382.570.351.520.760.230.536.200.2937.08

SE0.970.940.430.180.680.220.200.290.340.180.490.270.630.420.430.370.670.500.590.550.380.140.300.210.120.180.580.131.17

FemaledLPR32.9631.604.070.6212.351.730.992.352.960.376.051.984.20

1.731.601.6010.005.937.655.313.950.492.472.350.741.6011.360.3749.01

SE1.65***1.63***0.70**0.281.16***0.46*0.35*0.53**0.60*0.210.84***0.49**0.71***0.46**0.44**0.441.050.830.930.79*0.680.250.550.53***0.30*0.44**1.12***0.211.76***

Time2 (%)Male*

LP R16.3315.191.720.725.440.570.140.431.150.862.290.299.174.154.012.5811.756.598.747.883.300.721.290.140.000.146.450.2935.67

SE1.401.360.490.320.860.290.140.250.400.350.570.201.09

0.760.740.601.220.941.071.020.680.320.430.140.000.140.930.201.81

TotalfLP R25.2724.012.980.669.151.190.601.462.120.604.311.196.502.852.722.0610.816.238.166.503.650.601.921.330.400.939.080.3342.84

SE1.121.100.440.210.740.280.200.310.370.200.520.280.63

0.420.420.370.800.620.710.630.480.200.350.290.160.250.740.151.27

Note. DSM-IH-R = DiagnosticandStatistical ManualofMentalDisorders (rev.3rded.).a n = 8 9 1 . b=819. cn = l , 7 1 0 . d = 8 1 0 . e=698. f = l , 5 0 8 .* / > < . 0 5 . **p


7/12

ADOLESCENT PSYCHOPATHOLOGY 139

_^"-.II jCV^JSO

"S3 S-CS;5*,1QboSg5OiSIJO.0511iga

o 5a> -^IS "-1

o%Q

,ao

rt^"-IIC

^c3 itff iia;

upi .||^7 ?

8 m.J U s*^^~3 *~*W IIC

-yW U os ,-.fv j * vo mO -H f i1-* vo r*^ m ~ ' r j " " . X ~ ' < N " ~ ' ^ : d ^ Sr* t in

C u9 - .9 .9 -.9 .9 9 9 92 u g g 2 2 r t ^ ^ ^8 o & 8 .9 8 .9 8 .9 % 8 .9 8 .9 8 .9 8 .9IT- 11si1111i1 Ii1J1J1| S-| 1|1|3| S|S | S|Ii

1-seO3J2,J 0u^ st.eU0T3 oO2in-e1or"( -S"o^ S BoH Iv c3C *^ 3 ,c5MIICZ'C 30hts -Quog(5u ~ 011Is'J5J1 3il

v&2JgO"72 W


8/12

140 L E W IN S O H N , HOPS, ROBERTS, SEELEY, AN DR EW S

11QII

uItS~ Reap

irInd

* .sII

8I1

E

2

o

1

o

ra

0 3

a&

00

o:

Mo i

to

( 2

CO

3 .

O t---r--r--r-OOf oôor-oo orsooooo

O O O O O O O O O O c DO O O O O O O O O O O O O O O O O

TTTt OOOOOOO OOOOOOO Tt N NOOOOOOOO

NfNOO

OvO ^

t--OOOOOOOOr>OOOOOO tQON O O r - O O O O O O O O O O O ^ O f ^ O O O O OO O O -HO O

OOOOOO^QÔOÔO-Ô-ôooooooooooooooooocJoooooooo oo

OOOOOOfN~OO ' O O O P ^ O 1 ( N t X > V O V D r f ( N ( N ( N O p O < ^ O ( NOOOOOOOOOOOOOOOOOOOOOOOOOOOOTt^'OOOOOOOOOOOOOO-^'rJtN OOOOOO or^

--Oooooooooooooooooooooooooo ooor --(Nr~-o


9/12

ADOLESCENT PSYCHOPATHOLOGY 141trast, Time1pointa nd lifetimep revalence ratesfordisruptivebehaviordisorders fo rmale students were more than twiceaslarge as theratesforfemale students.AtTime1,older adolescents were m ore likelytohaveadiag-nosis of dysthymia (oddsratio [OR] = 1.87; 95%confidencebounds =1.03,3.40),and youn geradolescentswere m ore likelytohaveadiagnosisof adisruptive behavior disorder(O R=1.47;95% confidence bounds = 1.07, 2.01). Asexpected, age wasassociatedwiththelifetim eprevalence of mostofthedisorders,withthe excep tion of childhood d isorders, particularly disrup-tiveb ehavior disorders and some anxiety disorders (e.g., separa-tion anxiety, overanxious).The only significantage effect fo rthetotal incidencewa sfound for the omnibustesto fdevelop-ingany disorder: Older adolescents were mo relikelyto developadisorder between Times1 and 2 (OR = 1.19;95%confidencebounds= 1.03,1.37). No ageeffects fo rdisorder-specific inci-dence rates werefound.Th einteraction betweenage andgenderfortheoccurrenceofunipolar depression was notsignificant. Thus, thehypothesisthat the difference inoccurrence ofdepression between malean dfemalestude nts would increaseas afunctionof age was notsupported. However, significant Age X Gender interactionswerefound fo rTime1point prevalenceof theom nibus testo fany disorder (OR =1.32,95%confidence boun ds= 1.01,1.74)an d anxiety disorders(OR = 2.16; 95%confidence bounds=1.23,3.77)aswellas forTime 1lifetimeprevalenceof anxietydisorders (OR = 1.43; 95% confidence bounds = 1.06, 1.93).Post hoc comparisons revealed that femalestudents under16years of ageweresignificantlyhigher than male studentson thepoint prevalence of any disorder, x 2( l,N=613)=4.77,p


10/12

142 L E W I N S O H N , HOPS ROBERTS SEELEY, ANDREWSissuesw ehaveoutlined:Why is the ECAratesolow?On theother han d, why is our rate so high? The lowlifetimerates by theECAhave been recognized as problematic (Parker,1987;Rob-erts,1988),and it has been suggested that theDISmethodologyused in the ECA m ight have seriously underestimated lifetimerates. The ECAalsocontained the paradoxical finding of agradual diminu tion of thelifetimerate as afunctiono f age,witholder people ha ving lower rates than younge r people (Rob ins &Regier,1991).To resolve the discrepan cy between the ECA andou r lifetime rates fo r affective disorder, we suggest four hy-potheses:

1 . The prevalence of depression has been increasing and isno washighas indicated by the lifetimerates of the adolescentsinou rstudy.2. The pre valence of depression has always been a s high asindicated by our data, and the ECAdata grossly undere sti-mated the real lifetime prevalence of depression in adults andespecially in older people.3. Adolescents experience and report many relatively tran-sient and short-lived episodes that, although meeting DSM-HI-Rcriteria,are notrecalledo rreportedlaterin life.4. Our in terviewe rs systematically overdiagnosed depres-sion (i.e., their thresholds fo r determining the presence ofvariousdepression symptom swaslower than inother studies).Although it is difficult to rule ou t this last hypothesis, wethink that it isimplausible because(a) our interviewers rigor-ouslyappliedDSM -III-R criteria;(b) ourpo int prevalencefig-ure of 2.9 was within the range of values reported by otherstudies; (c) the m ean num ber of depression symptoms rated aspresentb y achild clinical psychiatrist (albeiton asmall num berofcases)was higher (5.8)than thoseof our interviewers (5.5;thisdifference was not significant); (d) the ratingsby thechildclinicalpsychiatrists who rated all cases of depression at Tim es1 and 2 indicated that they considered approximately 90% ofthem to be m oderately to severely depressed and in need oftreatment; and (e) ourpointprevalence for anydisorderof9.7%wa sat the low end of the range of values (6.6%-37%) that hasbeen reported and was close to the10%estimated by Schwartz-Gould et al. (1981). Whether the lifetime prevalence was ashigh as suggested by our study or as low as in the ECA hasimportant implications that can be resolved onlybyfuture re -search.Therearefewstudiesin theliterature that provide incidencedata.AsEatonet al.(1989)pointed out, incidence dataa rehardto come by because, ideally, such data should be based on alongitudinal, prospective study design. Becausethe incidenceof depression and of m ost other m ental disorders is low,such a study m ust start w ith a large numbe r of people who arenot in an episode at the beginnin g of an observation period inordertogenerateareasonablenumbero fincidence cases overa1-year period.Comparingou rannua lfirstincidence rateo f5.7%for MD Dwith that reported in the literature, we found that ours wassubstantially highe r than (a) the 1.6%reported by Eaton et al.(1989) for the ECA data; (b) the 0.52% reported by Hagnell,Essen-Moller, Lanke, Ojesjo, and Rorsman (1990) for theLundby study;and (c) the 0.23% reported byM urphy, Olivier,Monson,Sobol, andLeighton (1988). Althou gha longitudinal,prospective design is the metho d of choicefordeterm ining inci-dence,it can beestimated if thepoin t prevalenceand the meanepisodedurationareknown,asincidenceisequaltoprevalence

divided by mean duration (Kleinbaum, Kupper, & Morgen-stern,1982).Th eformuladetermineshowmany people havetodevelop the disorder overa period oftime to keepthe preva-lence rate constant. Fo rexample,if the point prevalence is 4%and the episode duration is 6 mon ths, then 8% of the popula-tion havetobecome depressedevery yearinordertokeep theprevalence at 4%. If weapply thisformula, weobtain an esti-mated incidencefor MDD of5.8%for theadolescents from thepresent study (the prevalence was 2.6% and mean duration was5.4 m onths). It thus seems tha t our 1-year incidence rates areclosetowhatcan be expected mathem atically.Whethero nelooksat first incidence, total incidence,or therelapse rate, our results suggest that the nu mb er of high schoolstudents who become depressed du ring a1-yearperiod ishigh.Projected on to a high school with1,000students, ourdataindi-cate that during a 1-year period, approximately 42 studentswouldbecome depressed fo rthe first time in the ir lives and thatamong those with a previous history of depression but whowere not depressed initially at Time 1(n = \74), 32 would be-come depressed again. Thus, in our hypothetical school, 74cases of depression would beexpected tooccur during the 1 -year period. The fact that the numbero fadolescents who be-come depressed overa 1-year period a pparently isthis largeha simportant implications for the need fo rprograms to detect,refer, an d treat thisag egroup.The degree of comorbidity between all of the major DSM-III-R disorders wasfound to be substantial (i.e., those withadisorder have a markedly elevated probability of also havinganother disorder). There have been previous reports on the Co-morbidityof major depression with dysthymia (Lewinsohnetal.,1991)and between u nipolar depression and o ther disorders(Rohdeet al.,1991)for theOADP sample.In the future wewillexamine the effects ofcomorbidity on psychosocial function-ing, suicidal behavior, treatment seeking, and other outcomemeasures.Having a previous episode of any disorder is a strong riskfactor fo r having another episode of mental disorder (OR =1.74;95%confidence bounds = 1.31, 2.31).Th e onlydisorderthat did no t seem to fit this pattern was adjustmen t disorder.However,datafor11of the 12subjects with a lifetime diagnosisof adjustmen tdisorder fo rwhomfollow-updata were availableindicated that 5 of them had developed an episode of majordepression between Times1 and 2.Thus, adjustment disorderis a strong riskfactor fo rmajordepression. Another importantfindingwas that there were no newcaseso fbipolar disorderand that all of the cases of bipolar disorder at Tim e 1 (most ofwhom were cyclothymic) had rem itted by Time 2.Th e effectof age on prevalencean d incidenceofdepressionwa sn otsignificant.Thus, our hypothesis of an increase in de-pression between 14 and 18years of age was notsupported.That there was no increase as a function of age over this agespan was consistent with recent findings on self-report mea-sures such as the CES-D and the BDI (Roberts et al., 1990,1991).A s expected, femalestudents scored substantially higherthan malestudentson all indexesofunipolar depression,whichwas consistent with other studies such as those by Kashani,Carlson, et al. (1987), McGee et al. (1990), and Kandel andDavies (1982). On the other hand, the critical interaction be-tweenage and genderon depression prevalencean d incidencewa s no t significant. In other words, female studentsscored


11/12

ADOLESCENT PSYCHOPATHOLOGY 143higheron thedepression measuresfor thetotalagerange thatwetested;the gapthat already existedat age 14betweenfemaleand male students did not widen. Thus, to find the age at whichfemale adolescents begintosurpass male adolescents willre-quirestudiesofearly adolescents.

Disruptivebehavior disorders were more frequent inmalestudents regardless of age. This was consistent with the findingsof many studies, such as those by Offord et al. (1987),whof oundthat male adolescentshad ahigher prevalenceofconductdisordersregardless of age. Similarly, thepoint prevalenceforsubstance use disorders was consistent with other studies(e.g.,Robins&Regier,1991)inshowing higher levelsbymale adoles-cents.For theTime1lifetime prevalenceand 1-yearincidenceresults, the trends were inthesame direction but were notsignif-icant.A sindicated earlier,it isimportant thatthehighlifetimeand1-yearincidence ratesweobtained forunipolar depression becross-validated in other studies. Pending cross-validation,wesuggest that our results can be generalized to adolescents inhighschoolinsmall-andmedium-sized urban areasofpredom-inantlymiddle- andupper-middle-classpeople.Thelong-term consequencesofhavinganepisodeofdepres-sion or another mental disorder during adolescence need to bestudied. By following our sample, wehope to provide moreinformationabout the course of episodes of disorders duringadolescence and to contribute to the understanding of the ante-cedentsand the consequences of mood and other mental dis-orders during adolescence,aswellas the implicationsfor de-pressioninlaterlife.

ReferencesAmer ican Psychiatric Association. (1980).Diagnostican dstatisticalmanualofmentaldisorders(3rded.).Washington,DC :Author.American Psychiatric Association. (1987).Diagnostic and statisticalmanualofmentaldisorders(Rev.3rded.).W ashington,DC:Author.Angold,A.(1988).Childhood and adolescentdepression:I. Epidemio-logical an d aetiological aspects. British Journalof Psychiatry,152,601-617.Angold, A.,Weissman, M. M., John,K.,Wickramaratne,P., &Prusoff,B.(1991).Theeffectsof age and sex on depression ratings in childrenan dadolescents.Journal o fth eAmerican AcademyofChild andAdo-lescentPsychiatry, 30 ,67-74.Angst,J., Dobler-Mikola,A.,& Binder, J. (1984). The Zur ichStudy Aprospective epidemiological studyof depressive, neurotic an dpsy-chosomatic syndromes: I. Problem, methodology. European Ar-chivesof Psychiatric NeurologicalSciences,234,13-20.Beck, A. T.(1967).Depression:C linical, experime ntal,and theoreticalaspects. Ne wYork:Harper &Row.Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J.(1961). An inventoryfo rm easuring depression. ArchivesofGeneralPsychiatry, 4,561-571.Bird, H. R.,Canino, G., Rubio-Stipec, M., Gould, M. S.,Ribera, J.,Sesman, M .,Wo odbury,M .,Huertas-Goldman,S.,Pagan, A.,San-chez-Lacay,A.,& Moscoso, M. (1988). Estimates of the prevalence ofchildhood maladjustment in a community survey inPuerto Rico:The use of combined measures. Archivesof GeneralPsychiatry, 45 ,1120-1126.Canino, G.J.,Bird, H. R., Shrout, P.E.,Rubio-Stipec, M., Bravo, M.,Martinez, R.,Sesman, M ., & Guevara, L. M.(1987).The prevalenceof specificpsychiatricdisorders in Puerto Rico. ArchivesofG eneralPsychiatry, 44 , 727-735.Carlson, G. A., &Strober, M . (1978). Affective disorders in adoles-

cence: Issues in misdiagnosis.Journalo fClinicalPsychiatry, 39 , 63-66.Carlson, G. A., &Strober, M . (1979). Affectivedisordersinadoles-cence. PsychiatricClinicsofNorth America,2 ,511-526.Chambers, W J.,Puig-Antich,J.,Hirsch, M., Paez, P.,Am brosini,P. J.,Tabrizi, M. A., & Davies, M. (1985). The assessment of affective

disordersinchildrenandadolescentsbysemistructured interview:Test-retest reliability of the Schedule for Affective Disorders andSchizophrenia for School-Age Childre n, Present Episode Version.ArchivesofGeneralPsychiatry, 42 ,696-702.

Chiles, C. A., Miller, M. L., &Cox,G. B.(1980). Epidemiology ofdepressive symptomatology in adolescence. Journalof AmericanAcademy ofChildPsychiatry, 23,91-98.Cytryn,L .,& M cKnew, D. H.(1972).Proposedclassification of child-hood depression.AmericanJournalof Psychiatry, 129,140-155.Eaton,W W ,Kram er, M., Anthony,J.C., Dryman, A.,Shapiro, S.,&Locke, B. Z. (1989). The inciden ce ofspecificDIS/DSM-IIImentaldisorders: Data from the NI M H Epidemiologic Catchment AreaProgram. A da Psychiatrica Scandinavica,79 ,163-178.Edelbrock, C., Costello, A.J.,Dulcan, M.K., Kalas, R.,& Conover,N. C. (1985).A gedifferences in the reliabilityof the psychiatricinterview of the child.Child Development,56, 265-275.

Endicott,J.,Cohen,J.,Nee, J.,Fleiss,J.,&Sarantakos,S.(1981).Hamil-ton Depression Rating Scale Extracted from regular and changeversions of the schedule foraffective disorders and schizophrenia.ArchivesofGeneralPsychiatry, 38 ,98-103.Endicott, J.,&Spitzer,R. L.(1978).Adiagnostic interview:The Sched-ule for Affective Disorders an d Schizophrenia. Archivesof GeneralPsychiatry, 35 ,837-844.Fleming, J. E., & Offord, D. R. (1990). Epidemiologyo f childhooddepressive disorders: Acritical review.JournaloftheAm erican Acad-em y ofChild andA dolescentPsychiatry, 29 ,571-580.Fleming, J. E., Offord, D. R., & Boyle, M. H. (1989). Prevalence ofchildhood and adolescent depression in the community: OntarioChildHealthStudy.British JournalofPsychiatry, 155,647-654.Hagnell, O, Essen-MOller, E., Lanke, J., Ojesjo, L., & Rorsman, B.(1990). Th e incidenceof mental illnessover a quarter of a century.Stockholm, Sweden: Almgrist&Wiksell.Hamilton, M .(1960).Arating scale fordepression. Journalo f 'Neurol-ogyand Neurosurgery, 23 ,56-61.Harrington, R., Fudge, H., Rutter, M ., Pickles, A.,&Hi l l ,J .(1990).Adultoutcomes of childhoo d and adolescent depression.ArchivesofGeneralPsychiatry, 47 ,465-473.Hirschfeld, R. M ., &Cross, C. K. (1982). Epidemiology of affectivedisorders. Archivesof GeneralPsychiatry, 39 ,35-46.Hops, H.,Lewinsohn,P. M.,Andrews,J.A .,&Roberts, R. E.(1990).Psychosocial correlates of depressive symptomatology among highschool students. JournalofClinical and ChildPsychology, 19 ,211-220.Kandel,D.B.,& Davies, M.(1982).Ep idemiology of depressive m oodin adolescents: An empirical study.Archivesof General Psychiatry,59,1205-1212.Kandel,D.B.,&Davies,M .(1986). Adult sequelaeofadolescent de-pressive symptoms.Archivesof GeneralPsychiatry, 43 ,255-262.Kashani ,J.H.,Beck,N.C., Hoep er, E. W,Fallahi,C,Corcoran, C.M.,McAllister,J.A.,Rosenberg,T .K.,&Reid,J. C.(1987).Psychiatricdisorders in a c om mu nity sample of adolescents. Am erican Journalo f Psychiatry,144,584-589.Kashani , J.H .,Carlson, G.A.,Beck,N.C., Hoeper, E. W, Corcoran,C. M ., McAllister, J. A., Fallahi, C., Rosensberg, T. K., & Reid,J.C.(1987). Depression, depressive symptoms,an d depressed moodam ong a community sample of adolescents. American JournalofPsychiatry,144,931-934.Kashani, J.H ., McGee, R. O.,Clarkson,S. E.A.,Walton, L.A .,Wil-liams, S., Silva, P. A.,Robins, A. J.,Cyt ryn,L., &McKnew, D. H.


12/12

144 LEWINSOHN, HOPS, ROBERTS, SEELEY,ANDREWS(1983). Depression in asample of9-yearo ldchildren: Prevalencean d associated characteristics. Archivesof G eneral Psychiatry, 40 ,1217-1223.Kazdin , A . E.(1989).Identifyingdepression inchildren:Acompari-son ofalternativeselection criteria.JournalofAbnormal Child Psy-chology,77,437-454.

Keller, M . B.,Lavori, P.W , Friedman, B.,Nielsen, E.,Endicott,J., &McDonald-Scott, P. A. (1987). TheLongitudinal IntervalFollow-UpEvaluation:Acomprehensive method fo rassessing outcome in pro-spective longitudinal studies. Archiveso f GeneralPsychiatry, 44 ,540-548.

Kleinbaum, D. G, Kupper, L. L., &Morgenstern,H. (1982). Epidemio-logic research.Belmet,C A :Lifetime Learn ing Publications.

Kovacs,M., &Beck,A. T.(1977).Thewishtodie and thewishtoliveinattempted suicides.Journalof ClinicalPsychology, 33 ,361-365.Kovacs, M.,Feinberg,T.L.,Crouse-Novack,M .A.,Paulauskas,S.L.,&Finkelstein,R.(1984).Depressive disorders inchildhood:I . Alongi-tudina l prospective studyofcharacteristics an d recovery.ArchivesofG eneral Psychiatry, 41 ,229-237.Lewinsohn, P. M, Rohde, P.,Seeley, J. R., & Hops, H. (1991). Th ecomorbidity ofunipolar depression: I.Major depression with dys-thymia.JournalofAbnormalPsychology, 100,205-213.McGee,R .,Feehan,M .,Williams,S.,Partridge,F, Silva,P.A.,&Kelly,J.(1990). DSM -IIIdisordersin alarge sampleofadolescents.Journal

of theAmerican Academy ofChild an d AdolescentPsychiatry, 29 ,611-619.Mitchell , J., McCauley,E., Burke,P.M ., &Moss,S. J. (1988). Pheno-menologyo fdepression inchildrena ndadolescents. Journalof th eAmerican AcademyofChildandAdolescentPsychiatry, 27 ,12-20.Murphy,J.M .,Olivier,D.C,Monson,R .R.,Sobol,A . M., &Leighton,A . H. (1988). Incidence of depression an d anxiety:T he StirlingCounty Study.American JournalofPublicHealth,78,534-540.Nat iona l Center fo r Health Statistics. (1989). Vital statistics of theUnited States:1987 . Deaths for selected causes (Tables 290, 292).Rockville, MD:Author.Newcomb, M. D, &Bentler, P. M. (1988).Impactofadolescent druguse andsocial support ofproblemso fyoung adults: Alongitudinalstudy.JournalofAbnormal Psychology, 97 ,64-75.Offord, D. R. (1985).Child psychiatric disorders: Prevalenceand per-spectives.PsychiatricClinicsof NorthAmerica,8 ,637-652.Offord, D. R.,Boyle,M .H.,Szatmari, P.,Rae-Grant, N. I.,Links,P.S.,Cadman, D. T, Byles,J.A .,Crawford,J . W,Blum,H.M .,Byrne, C,Thomas, H, &Woodward, C. A. (1987).Ontario Child HealthStudy:II. Sixmonth prevalenceofdisordera ndratesofservice utili-zation.Archivesof GeneralPsychiatry, 44 ,832-836.

Orvaschel,H.,Puig-Antich,J., Chambers,WJ.,Tabrizi,M.A.,&John-son, R. (1982). Retrospective assessment ofprepubertalmajorde-pression with th e Kiddie-SADS-E. Journalof th eAmerican Acad-em y of Child Psychiatry, 21 ,392-397.Parker,G(1987). Are the lifetime prevalence estimates in the EG AStudy accurate?PsychologicalMedicine,17 , 275-282.Paton,S.,Kessler,R.,&Kandel,D.(1977).Depressive moodandado-lescent drug-use:Alongitudinalanalysis.JournalofG eneticPsychol-ogy, 131,267-289.Pearson, R. W,Ross, M .,&Dawes, R. M.(1991).Personal recall an dthe limitsofretrospective questions insurveys.In J . M.Tanur (Ed.),Questionsabout surveyquestions:Inquiriesintothe cognitivebasesofsurveys (pp.65-94).New\brk:Sage Russell Foundation.Poznanski,E.O,&Z r u l l ,J. P.(1970).Childhooddepression.Archivesof GeneralPsychiatry, 23 ,8-15.Puig-Antich,J.,&Chambers, W J. (1983).Schedulefo r Affective Dis-ordersand Schizophrenia for School-Age Children 6-18years). Pitts-burgh:Western Psychiatric InstituteandClinic.

Radloff, L. S.(1977).T heCES-D Scale:Aself-report depression scalefor research in the general population. Applied PsychologicalMea-surement,1 ,385-401.Roberts,R . E. (1988). Theepidemiologyofdepressioninminorities.InP.Muehrer (Ed.),Researchperspectivesondepression and suicideinminorities(pp. 1-20). Washington,D C:National InstituteofMentalHealth.Roberts, R. E.,Andrews,J. A.,Lewinsohn,P.M.,&Hops,H. (1990).Assessmentofdepression inadolescents usingtheCenter fo rEpide-miologicStudies Depression Scale.JournalofC onsulting andClini-ca lPsychology, 2,122-128.Roberts, R. E.,Lewinsohn,P.M.,&Seeley,J. R. (1991).Screeningforadolescent depression: Acomparison ofdepressionscales.Journal

of th eAmerican Academy ofChild an d AdolescentPsychiatry, 30 ,58-66.Robins, L.,&Regier,D.(1991).PsychiatricdisordersinAmerica.N ewYork:Free Press.Robins ,J.N,Helzer,J.E.,Crougham,R.,Williams,J. B. W, &Spitzer,R. L. (1981).NIMH DiagnosticInterviewSchedule:VersionH I.Rock-ville, M D :National InstituteofMental Health.Rohde, P.,Lewinsohn,P .M ., &Seeley, J. R. (1991).Comorbidity ofunipolar depression:II.Comorbidity withothermentaldisordersinadolescents an d adults.JournalofAbnormal Psychology, 100,214-222.

Rutter,M. (1988).Epidemiological approaches todevelopmental psy-chopathology.Archivesof GeneralPsychiatry, 45 ,486-495.Rutter,M .,Schaffer, D.,&Shepherd,M . (1975).Amultiaxial classifica-tion of child psychiatric disorders. Geneva, Switzerland: WorldHealthOrganization.Rutter,M .,Tizard,J.,Yule,W ,Graham,P., &Whitmore,K.(1976).Isleof Wight studies, 1964-1974. PsychologicalMedicine,6, 313-332.Schwartz-Gould, M ., Wunsch-Hitzig,R., &Dohrenwend, B.(1981).Estimating the prevalenceo fchildhood psychopathology:Acriticalreview.Journal of the American Academy ofChild Psychiatry, 20 ,462-476.Shaffer,D.,Gould,M .S.,Brasic,J .,Ambrosini, P.,Fisher,P.,Bird,H.,&Aluwah l i a , S. (1983). Th e Children s Global Assessment Scale(CGAS).Archivesof GeneralPsychiatry, 40 , 1228-1231.Shapiro,R.,&Keller,M.(1979).Longitudinal Interval Follow-UpEval-uation (LIFE). Unpublished manuscript.Spitzer, R. L.,Endicott,J.,&Robins, E.(1978). Research DiagnosticCriteria (RDC).Psychopharmacology Bulletin,11,22-24.Teri,L.(1982).The use of theBeckDepressionInventory withadoles-cents.JournalofAbnormal ChildPsychology, 10 , 277-284.Velez,C.N. ,Johnson,J.,&Cohen,P .(1989).Alongitudinal analysisofselected risk factorsfo rchildhood psychopathology: Th eChildrenin the CommunityStudy.Journal American Academy ofChild an dAdolescent Psychiatry, 28 ,861-864.Weinberg,WA .,Rutman,J.,Sull ivan, L.,Penick,E .C.,&Dietz,S. G.(1973).Depression inchildren referred to aneducationaldiagnosticcenter: Diagnosis an d treatment. Journal of Pediatrics, 83, 1065-1072.

Weissman, M. M., & Myers,J. K. (1978).Affective disorders in aU.S.urban community.Archivesof GeneralPsychiatry, 35 ,1304-1311.Whitaker, A. , Johnson, 1, Shaffer, D, Rapoport,J. L., Kalikow,K. ,Walsh,B. X,Davies,M .,Braiman,S.,&Dolinsky,A.(1990).Uncom-montroublesinyoungpeople:Prevalenceestimatesofselectedpsy-chiatric disordersin anonreferredadolescentpopulation.ArchivesofGeneral Psychiatry, 47 ,487-496.

Received July8 ,1991Revision received June25,1992Accepted July10,1992

Depression Adolescence

Documents